Cyclic ADP-ribose Enhances Coupling between Voltage-gated Ca2+ Entry and Intracellular Ca2+ Release*

Ca2+ release from intracellular stores can be activated in neurons by influx of Ca2+through voltage-gated Ca2+ channels. This process, called Ca2+-induced Ca2+ release, relies on the properties of the ryanodine receptor and represents a mechanism by which Ca2+ influx during neuronal activity can be amplified into large intracellular Ca2+ signals. In a differentiated neuroblastoma cell line, we show that caffeine, a pharmacological activator of the ryanodine receptor, released Ca2+ from intracellular stores in a Ca2+-dependent and ryanodine-sensitive manner. The pyridine nucleotide, cyclic ADP-ribose, thought to be an endogenous modulator of ryanodine receptors also amplified Ca2+-induced Ca2+ release in these neurons. Cyclic ADP-ribose enhanced the total cytoplasmic Ca2+ levels during controlled Ca2+ influx through voltage gated channels, in a concentration-dependent and ryanodine-sensitive manner and also increased the sensitivity with which a small amount of Ca2+ influx could trigger additional release from the ryanodine-sensitive intracellular Ca2+ stores. Single cell imaging showed that following the Ca2+ influx, cyclic ADP-ribose enhanced the spatial spread of the Ca2+ signal from the edge of the cell into its center. These powerful actions suggest a role for cyclic ADP-ribose in the functional coupling of neuronal depolarization, Ca2+entry, and global intracellular Ca2+ signaling.

neuronal activity (13), during changes in synaptic efficacy that may be involved in learning processes (14 -16), as part of the mechanism of neurotransmitter release (17), and even during neuronal development (18). Caffeine (4) and cyclic ADP-ribose, the novel Ca 2ϩ -mobilizing pyridine nucleotide originally discovered in the sea urchin egg (19), both release Ca 2ϩ from intracellular stores via modulation of the ryanodine receptor (20,21). In the last few years cyclic ADP-ribose has begun to emerge as a potential physiological regulator of ryanodinesensitive Ca 2ϩ -dependent processes in a number of intact mammalian systems. Cyclic ADP-ribose modulates excitationcontraction coupling in the heart (22), it alters excitability of pancreatic acinar cells (23) and dorsal root ganglion cells (24), and stimulates Ca 2ϩ release from the intracellular stores of T-lymphocytes (25). Using a combination of electrophysiology and Ca 2ϩ imaging we show here, in intact mammalian cultured neuroblastoma cells (26), that release from ryanodine-sensitive intracellular stores is coupled to Ca 2ϩ influx via voltage-activated channels and potentiated by cyclic ADP-ribose applied through the recording electrode.
EXPERIMENTAL PROCEDURES NG108-15 neurons, a mouse neuroblastoma ϫ rat glioma hybrid culture (obtained from the European Collection of Cell Cultures, Porton, UK), were cultured as described previously (26,27).
For Ca 2ϩ measurements in intact cells, the cells were loaded with Fura-2 using the acetoxymethyl ester loading technique for a maximum loading period of 15 min. Before commencing experiments the cells were washed three times in the perfusion buffer containing 130 mM NaCl, 10 mM HEPES, 5 mM KCl, 5 mM CaCl 2 , 1 mM MgCl 2 , 25 mM glucose, on the stage of the upright microscope (Zeiss, Oberkochen, Germany). To depolarize the cells, 30 mM KCl (and an equimolar reduction in Na ϩ ions) was switched into the perfusion flow (rate varied between 1 and 1.5 ml/min). Changes in Ca 2ϩ were measured every 4 s in these experiments using ratiometric determinations of image intensity following excitation with 340-and 380-nm wavelength light supplied by a TILL Photonics monochromator (Planegg, Germany) controlled by Improvision (Leicester, UK) "Ionvision" software (as described previously (28)). An in vitro calibration using the Ionvision software was used to analyze average Ca 2ϩ changes, over the whole cell, in individual cells from the pseudocolor images, as described previously. The same experimental set up and analysis was used to determine Ca 2ϩ changes during the electrophysiological experiments, except that an image was captured every 0.8 -1.2 s. Variations in intracellular Ca 2ϩ in different regions of the cell were also measured with the same calibrations and software using small rectangular regions of interest. Electrophysiological recordings were made in the whole cell patch clamp mode using an Axopatch 200A (Axon Instruments, Foster City, CA) with pipettes of resistance 2-6 M⍀. Seal resistances prior to breakthrough were always greater than 1 G⍀. Cells settled and filled with Fura-K ϩ for approximately 10 min after breakthrough. Current and voltage signals were digitized using an ITC-16 A/D converter (Instrutech Corp., Great Neck, NY) and the experiments controlled using the Axodata program (Axon Instruments) through the same interface. The intracellular solution contained 135 mM CsCl, 10 mM HEPES, 1 mM Mg-ATP, 100 M Fura-2-pentapotassium salt, and extracellular solution contained 130 mM NaCl, 10 mM HEPES, 5 mM KCl, 5 mM tetraethylammonium chloride, 5 mM CaCl 2 , 5 mM 4-aminopyridine, 1 mM MgCl 2 , 25 mM glucose, and 1 M tetrodotoxin. Both intra-and extracellular solutions were designed to block K ϩ and Na ϩ channels, so that during cellular depolarization the area beneath the inward current, measured with Axograph (Axon Instruments), an indication of the charge entering the cell, represented the entry of Ca 2ϩ and not other cations. Ca 2ϩ (or charge) entry was controlled by changing the duration of the ϩ60or ϩ80-mV voltage step evoked from a holding potential of Ϫ70 or Ϫ90 mV. Voltage steps were applied in a random order, every 20 -30 s to avoid excessive run-down of the Ca 2ϩ current with linear on-line leak subtraction. For each voltage step the peak Ca 2ϩ change was measured and divided by the * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. area beneath the current trace (pA ϫ s or pC) to express the unit Ca 2ϩ transient (29), Ca 2ϩ /pC (pA ϫ s) charge entering the cell. Three voltage steps were used to calculate a mean value of the unit Ca 2ϩ transient at each step duration. All experiments were conducted at room temperature. All values are compared with a Student's t test, and values are means Ϯ standard error of the mean. All materials were obtained from Sigma (Poole, UK) except Fura-2, which was from Molecular Probes.

RESULTS AND DISCUSSION
Initial studies showed that these differentiated neuroblastoma cells responded to high concentrations of caffeine, 20 -50 mM, with small increases in intracellular Ca 2ϩ (Fig. 1A). The responses persisted in the absence of extracellular Ca 2ϩ , indicating that these cells possess an intracellular, caffeine-sensitive Ca 2ϩ store. If the NG108-15 cells were first depolarized with 30 mM K ϩ , we observed a rise in intracellular Ca 2ϩ as observed previously (30), consistent with influx of Ca 2ϩ through N and L-type voltage-sensitive Ca 2ϩ channels present on the plasma membrane (open bar, Fig. 1B). Application of caffeine, immediately after the depolarization, then gave a fast and large intracellular Ca 2ϩ rise ( Fig. 1, B and C). The responses resembled those seen in bullfrog and rat sympathetic neurons and also in rodent central neurons (6,7,31). The caffeine responses were blocked by ryanodine (5-10 M), an antagonist of the ryanodine receptor at these concentrations (see legend to Fig. 1). This result shows that Ca 2ϩ entry by prior depolarization sensitized the ryanodine receptors on the intracellular stores to subsequent activation by caffeine and represented a form of CICR. A concentration response curve (Fig. 1D) shows a steep response to caffeine in the presence of the prior depolarization, strongly indicating the operation of an amplification process.
Having established that these neurons possessed a ryanodine-sensitive CICR capability, we next sought to estabish whether cADPR, a positive modulator of the ryanodine receptor, could potentiate CICR in these cells. Since cADPR is not membrane-permeable, we applied it to the cells via a whole cell patch pipette and simultaneously used voltage clamp to control the membrane potential of the cell. This allowed us to control Ca 2ϩ influx through voltage-sensitive Ca 2ϩ channels. Previous studies have successfully used a method called the unit Ca 2ϩ transient to relate the change in intracellular Ca 2ϩ levels directly to the amount of charge entering the cell (29, 32, 33). The amount of charge entering the cell was estimated from the area beneath the current trace that represented Ca 2ϩ entry, since K ϩ and Na ϩ channels were blocked. By dividing the peak Ca 2ϩ rise simultaneously recorded during the voltage step, by the area beneath the current trace, we calculated a unit Ca 2ϩ transient expressed as nanomolar Ca 2ϩ release per picocoulomb of charge entry. As shown in Fig. 2A, increasing the length of the voltage step from 100 to 1000 ms allowed the Ca 2ϩ channels to stay open for longer, thus allowing more Ca 2ϩ to enter the neuron. The cell in Fig. 2 was recorded using a pipette containing 10 M cADPR, and it can be seen that relative to a rather modest increase in the area beneath the Ca 2ϩ current traces, the longer voltage steps evoked a large increase in the intracellular, simultaneously measured Ca 2ϩ level (Fig. 2B). This was consistent with extra Ca 2ϩ being released from the intracellular stores triggered by the initial Ca 2ϩ entry and had the effect of increasing the absolute value of the unit Ca 2ϩ transient, above that seen in a control cell (Fig. 2C). Note also that the red Ca 2ϩ trace (Fig. 2B) showed an additional later Ca 2ϩ release following the initial peak, a common occurrence in cells treated with cADPR. The effects of cADPR present in the patch pipette are illustrated by the images of two cells with similar charge entry during a depolarizing pulse (Fig. 2D). Particularly notable is the larger Ca 2ϩ rise in the cADPRtreated cell compared with control. Also apparent, after application of the depolarizing pulse, was the rise in Ca 2ϩ at the edge of the cell before Ca 2ϩ rose in the center of the cell, and a striking difference was the relatively larger elevation in Ca 2ϩ at the center of the cADPR-treated cells compared with control. By placing two small regions of interest over the edge and center of the cells and using the Ca 2ϩ rises in these regions to calculate a unit Ca 2ϩ transient, we observed a 120% increase in the unit Ca 2ϩ transient in the center of cADPR-treated cells compared with the center of control cells (n ϭ 9). This suggested that cADPR increased the likelihood with which elevated Ca 2ϩ levels close to the plasma membrane propagated to the center of the cell to give a global Ca 2ϩ signal over the whole of the cell (34).
When Ca 2ϩ levels over the whole cell were measured and compared for all cells, as the duration of the voltage step was increased, the increase in the unit transient was greatest in the   FIG. 1. A, intracellular Ca 2ϩ levels were increased by 50 mM caffeine, indicated by the filled bar. B, depolarization of the cells with 30 mM external K ϩ , as shown by the open bar, lead to an increase in Ca 2ϩ levels through the opening of N and Ltype voltage-sensitive Ca 2ϩ channels (30). Subsequent application of 50 mM caffeine (filled bar) leads to a marked increase in Ca 2ϩ levels, as shown for all cells in C. In the absence of extracellular Ca 2ϩ , as shown by the dotted line, depolarizations did not lead to significant rises in intracellular Ca 2ϩ , nor did it enhance caffeineinduced Ca 2ϩ release, rather Ca 2ϩ changes were not significantly different to the results when caffeine was applied alone, 172 Ϯ 20 nM Ca 2ϩ (n ϭ 5) p ϭ 0.24, t test. These responses were reduced by extracellular application of 10 M ryanodine, mean caffeine-induced Ca 2ϩ rises were 658 Ϯ 113 reduced to 91 Ϯ 25 nM Ca 2ϩ (n ϭ 8, p ϭ 0.004, paired t test). D, concentration response curve showing the sharp increase in the predepolarization caffeine-induced intracellular Ca 2ϩ rises in response to raised caffeine concentrations, values are means Ϯ S.E. of the mean for at least three separate experiments and between five and nine cells. cells treated with cADPR (Fig. 3A). Controls also demonstrated a rise in the unit transient (33) as more Ca 2ϩ entered the cell, and it is tempting to suggest that this was due to low levels of endogenous cADPR found in a number of brain preparations (35). In response to the short 100-ms duration voltage step, the unit Ca 2ϩ transient was approximately 1.4 when charge (or Ca 2ϩ ) entry was 25.4 Ϯ 3.6 pC, rising to a value of around 2.5 (see Fig. 3A), indicating additional release of Ca 2ϩ from the internal stores, as the amount of charge (or Ca 2ϩ ) entry was increased to 101.6 Ϯ 14.5 pC as the length of the voltage step increased to 1000 ms. Addition of 10 M cADPR to the patch pipette increased the unit Ca 2ϩ transient in two ways. First it increased the value of the unit Ca 2ϩ transient, compared with control, even following a short, 100-ms duration voltage step (Fig. 3, filled circles compared with open squares, p ϭ 0.007, t  test). A direct comparison shows that for a similar charge (or Ca 2ϩ ) entry to the controls, 30.7 Ϯ 11.6 pC (p ϭ 0.24, t test), during the shortest 100-ms duration voltage step, the presence of cADPR in the pipette gave rise to a unit Ca 2ϩ transient of approximately 2.8, consistent with release of Ca 2ϩ from intracellular stores. Since a similar charge entry in controls failed to elicit Ca 2ϩ release from the intracellular stores, but 100 pC could, our result indicates that cADPR reduced, by approximately 3-fold, the amount of Ca 2ϩ entry required to trigger additional Ca 2ϩ release from the internal stores. Second, as the length of the voltage step was increased and more Ca 2ϩ entered the cell, the presence of cADPR resulted in a relatively larger increase in the unit Ca 2ϩ transient, compared with the controls (even though the absolute values of charge entry were similar between control and cADPR-treated cells, p ϭ 0.34, t test). This suggested that cADPR allowed the extra Ca 2ϩ entry evoked by depolarizations longer than 100 ms, to trigger even further release of Ca 2ϩ from the intracellular stores. We noted that the values of the unit transient were similar to those observed in isolated dorsal root ganglion cells (33,36), but approximately 10 ϫ larger than those observed in bullfrog neurons (29). This difference could relate to the much larger amplitude of the Ca 2ϩ currents in the bullfrog neurons and may reflect a greater degree of Ca 2ϩ buffering of the larger Ca 2ϩ entry compared with the smaller currents evoked in these mammalian neurons.
Inositol trisphosphate also enhances release of Ca 2ϩ from intracellular stores in a Ca 2ϩ -dependent manner, so called inositol trisphosphate-induced Ca 2ϩ release (IICR) (2,34). However we did not observe any effect on the unit Ca 2ϩ transient when 50 M InsP 3 was included in the patch pipette, even though a number of studies suggest that these neuroblastoma cells express InsP 3 -sensitive intracellular Ca 2ϩ stores (37-39) (mean value in the presence of InsP 3 following a 1-s pulse was  6). This confirms previous studies, which also suggest that the initial fast inactivating current is a result of the activation of an N-type channel (41,42), while the later component results from the opening of an L-type Ca 2ϩ channel (42). B, below the current records are the simultaneously measured Ca 2ϩ levels inside the cell. The colors of the Ca 2ϩ traces correspond to the colors of the current traces. The arrow shows the time the voltage step was applied, and note the different time scales of the Ca 2ϩ changes and the current responses. The pipette used to record this cell contained 10 M cADPR that potentiated the measured Ca 2ϩ rises when longer voltage steps, leading to prolonged inward currents (red, green, and yellow) triggered an enhanced Ca 2ϩ release. By measuring the area beneath the current trace, pA multiplied by seconds, we obtained a direct indication of the charge, pC, entering the cell. Since all channels except voltage-sensitive Ca 2ϩ channels were pharmacologically blocked, this charge entry represented Ca 2ϩ entry. This charge was then expressed as a function of the Ca 2ϩ rise measured inside the cell, to give the unit Ca 2ϩ transient. In C, the unit transient calculated for this cell at each duration voltage step (from three separate voltage step applications) is plotted against the duration of the step. Also shown, in black, are values from a representative control cell. D, pseudocolor images of a control cell and a cell treated with 10 M cADPR. Images are sequential left to right and taken at 1.4-s intervals, the white arrow represents the time at which the voltage step was applied. The charge entry was similar in both cells. The pseudocolor images show that the edges of the cells were always the first regions to increase, presumably as Ca 2ϩ entered the cell following the depolarization. In the cADPR-treated cell there was a significant rise of Ca 2ϩ in the center of the cell after the influx, whereas in the control cell the rise in Ca 2ϩ in the center was less apparent and more confined to the edges. Calibration of the pseudocolor is shown and ranges from 0 to 1400 nM Ca 2ϩ , blue to red. The pipette is attached to the cell at the top right-hand corner of each cell. Mean whole cell resistance was 783 Ϯ 70.3 M⍀ (n ϭ 38), and there was no significant difference in the whole cell input resistance between control cells and cells treated with cADPR (p ϭ 0.5, t test) or in the areas beneath the evoked Ca 2ϩ currents, in cADPR-treated cells compared with controls (p ϭ 0.34, t test).
10 M ryanodine, an inhibitor of CICR, decreased the unit Ca 2ϩ transient during the shorter, 100-ms voltage step and also blocked the increase in the unit Ca 2ϩ transient with increasing pulse duration in both control cells and in cells treated with 10 M cADPR (Fig. 3, open triangles) A similar blockade occurred when either 10 mM Mg 2ϩ or 20 M ruthenium red were included in the pipette solution (see legend to Fig. 3A). All these treatments indicate that exogenously applied cADPR, acting on or close to the ryanodine receptor, increased the sensitivity of CICR in these neurons. cADPR did this in two ways: first, it reduced the amount of Ca 2ϩ entry required to trigger additional Ca 2ϩ release from the internal stores and second, by increasing the value of the unit Ca 2ϩ transient, it increased the total amount of Ca 2ϩ released from intracellular stores for a given controlled Ca 2ϩ influx. This increase in the unit Ca 2ϩ transient occurred in a concentration-dependent manner (Fig. 3B). 50 and 100 M cADPR increased the unit Ca 2ϩ transient to a level 3-fold over control, indicating that cADPR action reached a maximum.
Our results show that this neuronal cell line possessed ryanodine-and caffeine-sensitive intracellular Ca 2ϩ stores. Cyclic ADP-ribose, in the presence of Ca 2ϩ influx, increased the amount of Ca 2ϩ released from the intracellular store in these cells and also increased the sensitivity with which Ca 2ϩ entry triggered additional Ca 2ϩ release leading to a global intracellular Ca 2ϩ rise. These actions of cADPR suggest a requirement for powerful regulatory mechanisms to control the production of cytosolic levels of cADPR from its ubiquitous precursor, ␤-NAD ϩ (40). As Ca 2ϩ release from internal stores becomes more widely recognized as part of neuronal Ca 2ϩ homeostasis, an increased understanding of the factors controlling endogenous levels of cADPR in neurons is now required. The absolute values of the mean unit Ca 2ϩ transient in the cADPR-treated cells were significantly increased at all pulse lengths compared with control (p Ͻ 0.05, t test, all duration pulses, n ϭ 5), and a large increase was seen as the pulse length increased. In the control cells the longer pulse lengths also evoked an increase in the unit Ca 2ϩ transient consistent with an extra release of Ca 2ϩ from the intracellular stores (n ϭ 7 cells, p Ͻ 0.05, t test, comparing 100 ms pulse duration with 1000-ms pulse duration). Note also that the unit Ca 2ϩ transient evoked by a 100-ms duration pulse in the cADPR-treated cells was significantly increased compared with the control, p Ͻ 0.05, t test. There were no differences in the basal Ca 2ϩ levels observed in control cells compared with those treated with 10 M cADPR, mean values were 278 Ϯ 24 nM for controls and 277 Ϯ 25 nM for cADPR-treated cells, p ϭ 0.48, t test. 10 M ryanodine reduced the change in the unit Ca 2ϩ transients in both control and cADPRtreated cells (n ϭ 3 for both, there being no significant difference in the effects of ryanodine in the two conditions, p ϭ 0.33, t test). Unit Ca 2ϩ transients at 1000 ms were also effectively reduced by the inclusion of 20 M ruthenium red into the pipette and also in separate experiments by 10 mM Mg 2ϩ (43), mean values were 1.5 Ϯ 0.5 and 0.8 Ϯ 0.8, n ϭ 4 and 3, respectively. B, concentration response curve for the actions of cADPR, showing the increase in the value of the unit Ca 2ϩ transient calculated during a 1000-ms voltage step, with different concentrations of intracellularly applied exogenous cADPR. Values are taken from between 5 and 8 cells at each concentration and represent a larger data set than shown in Fig. 3A.
Voltage-gated Ca 2ϩ Entry and Intracellular Ca 2ϩ Release 20970